Neonatal seizures are a frequent problem encountered in neonatal nurseries, but their significance is controversial. Some investigators regard newborn seizures as simply epiphenomena and reflective of brain injury while others note associated metabolic and physiologic aberrations implying that seizures per se are injurious to the central nervous system. Despite the controversy, current standards are such that seizures are treated, but the choice of anticonvulsants has not been on the documented superiority of one agent over another. Small numbers of patients treated without EEG confirmation or careful attention to toxicity have formed the rationale for treating neonatal seizures. Several treatment regimens have been advanced but the two drugs most commonly chosen are phenobarbital and phenytoin. We propose to evaluate the relative efficacies of phenobarbital and phenytoin in the treatment of neonatal seizures utilizing EEG criteria for the diagnosis and response to therapy, but also recording and charting the course of clinically noted abnormal movements. This study will achieve free levels of 25 mg/L for phenobarbital and 3mg/L for phenytoin. The study will be performed over five years in a single institution enrolling twenty infants yearly. Fifty infants in each group will be randomized to phenobarbital or phenytoin therapy. Because of marked variations in the binding of phenobarbital and phenytoin discovered in preliminary studies, in vitro binding analysis performed prior determination of plasma free and bound levels will allow a judgement of whether desired therapy will be judged at 60 minutes. If no adjustment is necessary, response to therapy will be judged at 60 minutes, and at 120 minutes if a dose adjustment is required. Continuous EEG monitoring will be performed for the first 24 hours of the study. The aim of therapy will be complete cessation of seizures, but following optimum use of these anticonvulsants a scale will be used to assess graded differences in therapeutic response to the drugs. Respiratory and cardiovascular parameters of toxicity will be studied. Alterations of plasma concentrations of factors known to affect binding will be studied. This study will be the first to assess the efficacy of the two most common drugs used in the treatment of neonatal seizures. The data obtained will have important implications for treatment regimens and form a basis for the subsequent evaluations of agents used in treating neonatal seizures.